ABSTRACT
Data guiding the duration and route of streptococcal bloodstream infection (BSI) treatment are lacking. We conducted a retrospective cohort study of adults hospitalized with uncomplicated streptococcal BSI in a large integrated healthcare system from 2013 to 2020. The exposures of interest were antibiotic duration (5–10 days vs. 11–15 days) and antibiotic route (oral switch vs. entirely intravenous). The primary outcome was a composite 90-day outcome comprised of all-cause mortality, recurrent streptococcal BSI, or readmission. We performed non-inferiority analyses for each exposure. Separate multivariable Cox proportional hazards regression models were constructed for each exposure. The antibiotic duration analysis included 1,407 patients (5–10 days, n = 246; 11–15 days, n = 1,161). We found that 5–10-day courses were non-inferior to 11–15-day courses (P = 0.047). The antibiotic route analysis included 1,461 patients (oral switch, n = 1,112; entirely intravenous, n = 349). Oral step-down therapy did not meet the criteria for non-inferiority (P = 0.06). In the adjusted models, no significant difference was found in the primary outcome rate by antibiotic duration or antibiotic route at discharge. We found that 5–10-day courses were non-inferior to longer courses, and thus may be a safe and effective treatment option in the treatment of uncomplicated streptococcal bacteremia. Randomized controlled trials are needed to confirm the equivalent outcomes with shorter regimens and to definitively determine the optimal antibiotic route on discharge.
KEYWORDS: Streptococcus, bloodstream infection (BSI), treatment, duration, oral switch
INTRODUCTION
Streptococcal bloodstream infections (BSIs) are common and associated with high morbidity and mortality. In the United States, more than 500,000 cases of BSI annually were associated with approximately 80,000 deaths, and streptococcal species are among the most frequent causative organisms (1, 2). Data guiding optimal treatment of bacteremia are limited, and consequently, there is wide variation in the duration and route of therapy (3–7).
This variability in practice represents a key opportunity for improving the treatment of streptococcal BSI. Unnecessarily prolonged antibiotic courses can cause antimicrobial resistance, drug toxicity, and Clostridioides difficile infections (8–13). In contrast to oral therapy, outpatient intravenous antibiotic therapy upon hospital discharge via a peripherally inserted central catheter carries the risk of central line-associated bloodstream infections and thromboses and may also incur excess cost to healthcare systems (14–18).
Efforts to mitigate the risks of antibiotic treatment must be balanced with the need for curative therapy. Recent landmark studies have shifted the paradigm in the treatment of several common infection syndromes by demonstrating non-inferiority of shorter antibiotic courses and oral antibiotic courses in comparison to longer courses and intravenous courses, respectively (19–21). In Gram-negative BSI, it has recently been established that outcomes are similar with 7-day courses versus more prolonged courses and with oral switch therapy versus entirely intravenous therapy (4, 22, 23).
Due to minimal data for the treatment of streptococcal BSI, there is a wide range in treatment duration, and courses are frequently longer than 10 days (24–26). Both entirely intravenous courses and courses that include oral switch therapy are commonly used (27–32). We aimed to determine if shorter antibiotic courses and oral switch treatment courses are non-inferior to prolonged and entirely intravenous courses, respectively, by conducting a retrospective cohort study in a large integrated healthcare system.
MATERIALS AND METHODS
Study design and patient population
We completed a retrospective cohort study of patients admitted to Kaiser Permanente Northern California (KPNC) hospitals with streptococcal bacteremia from 1 October 2013 to 30 September 2020. Eligible patients were ≥18 years old with streptococcal species isolated from one or more blood cultures collected during the study period. Additionally, eligible patients were hospitalized in association with positive blood culture(s) and discharged to home, home with home health services, or against medical advice. Patients discharged to a skilled nursing facility were excluded due to inability to reliably characterize antibiotic regimens. Exclusion criteria included complicated streptococcal bacteremia, which was defined as a concurrent diagnosis of any of the following conditions: endocarditis, osteomyelitis, spondylodiscitis, epidural abscess, meningitis, septic arthritis, prosthetic joint infection, liver abscess, and cardiac implantable electronic device infection. Furthermore, patients receiving >16 days of treatment were excluded as prolonged treatment courses may indicate that complicated streptococcal bacteremia was suspected clinically. In addition, patients receiving <5 days of anti-streptococcal antibiotic therapy were excluded to avoid including patients whose blood culture isolates were clinically suspected to represent blood culture contamination. Patients with a history of streptococcal bacteremia within 1 year before the index blood culture, polymicrobial bacteremia, hospice care at the time of discharge, and those who were not discharged on antibiotic therapy were excluded. Antibiotics were required on discharge to both assess outcomes during clinically comparable time points and to improve antibiotic therapy characterization. Also, patients who met the primary outcome within 10 days of initiating antibiotic therapy were excluded from the analysis, comparing 5–10-day versus 11–15-day treatment courses to allow for treatment group assignment before assessing for the outcomes. To ensure adequate characterization, KPNC enrollment was required for 1 year before and for 90 days after the index bacteremia, and KPNC drug coverage was required.
Data collection and validation
Microbiologic data, baseline characteristics, and patient outcomes were collected from the KPNC electronic health record (EHR). HIV status was ascertained from the KPNC HIV Registry. Natural language processing (NLP) was used to identify streptococcal isolates from positive blood culture results within the EHR. Antibiotic treatment data were collected from KPNC pharmacy databases. To ensure accuracy of cohort identification, we performed validation of randomized data sets representing 5% of exposure assignments and 5% of outcome determinations. The accuracy rate overall was high with an accuracy rate of 97.2% for assigning duration exposure, 98.6% for assigning route exposure, and 100% in determining outcome.
Statistical analysis
Exposures
The first exposure assessed was the duration of antibiotic therapy. Patients were categorized into either the 5–10-or 11–15-day treatment groups based on the total duration of antibiotic therapy. The second exposure assessed was the route of antibiotic therapy. Patients were assigned to the entirely intravenous group if they were discharged on intravenous antibiotic therapy, including if they were discharged on both intravenous and oral therapy, whereas those discharged on an oral-only antibiotic regimen were assigned to the oral switch group.
Outcomes
Outcomes were assessed within 90 days of hospital discharge. The primary outcome was a composite outcome comprised of all-cause mortality and clinical failure, i.e., recurrent streptococcal bacteremia or all-cause hospital re-admission. Secondary outcomes were recurrent streptococcal bacteremia, all-cause mortality, and C. difficile infection.
Analysis
For each exposure assessed, we described the demographic and clinical characteristics of each study group (i.e., 5–10- vs. 11–15-day treatment group, oral switch vs. entirely intravenous group) using counts/percentages for categorical variables and medians with interquartile ranges for continuous or discrete variables. We examined statistically significant differences between the study groups using χ tests or Fisher’s exact tests for categorical variables and Student’s t-tests or Kruskal–Wallis tests for continuous variables.
We used the Farrington–Manning method to test the hypotheses that 5–10-day treatment courses are non-inferior to 11–15-day courses, and that oral switch courses are non-inferior to entirely intravenous courses with respect to the primary composite outcome. A 5% margin of non-inferiority was used in both tests. A 10% margin of non-inferiority has been recommended in evaluating treatments for complicated urinary tract infections (UTIs), hospital-acquired pneumonias, and ventilator-associated pneumonias (33, 34). A more conservative margin of 5% was chosen due to the potentially severe consequences of treatment failure in streptococcal bacteremia.
Additionally, survival curves were calculated using the Kaplan–Meier method to estimate the risk of the composite outcome within 90 days of hospital discharge. Cox proportional hazards regression was used to model the association between antibiotic route/duration and risk of the composite outcome. Per our pre-specified analysis plan, variables were initially included in the multivariable model as potential predictors if they were significant at the 0.20 level in bivariate models. Variables significant at the 0.05 level were kept in the multivariable model. The multivariable models for each exposure were adjusted for BMI, Charlson comorbidity index, white blood cell count, discharge venue, use of immunosuppressive therapy within the preceding year, and streptococcal species.
RESULTS
A total of 9,493 patients were identified as having streptococcal bacteremia during the study period from 1 October 2013 to 30 September 2020. The most common reasons for exclusion were discharge to a venue other than home (e.g., discharge to a skilled nursing facility, a nursing home, hospice, an assisted living facility or death prior to discharge) (n = 2,761) (Fig. 1) and insufficient KPNC enrollment during the follow-up period (n = 2,190), The final eligible cohort comprised 1,461 patients. Antibiotic susceptibilities for the index streptococcal blood culture isolates from the final eligible cohort are summarized in Table S1.
Fig 1.
Flow diagram of cohort assembly.
Duration of antibiotic therapy (i.e., 5–10 vs. 11–15 days)
In the analysis comparing 5–10-day vs. 11–15-day treatment courses, 54 patients were excluded from the primary analysis due to meeting the primary outcome within 10 days of initiating antibiotic therapy. Of the 1,407 patients included in the analysis, 1,161 and 246 patients received 11–15 and 5–10 days of therapy, respectively.
Patient characteristics
Demographic and baseline characteristics are shown in Table 1. Overall, the median age was 68.8 years (IQR, 58.1–79.3), and the majority were men (54.4%). Patients in the 5–10-day treatment group were less likely to have a concurrent diagnosis of cellulitis (19.9% vs. 29.9%) but were more likely to have a concurrent diagnosis of pneumonia (54.1% vs. 45.7%). Patients were also more likely to be dialysis dependent in the 5–10-day treatment course group (5.3% vs. 2.1%). There were differences noted in the streptococcal species isolated, for example Streptococcus pyogenes was less frequently isolated among patients in the 5–10-day treatment group (6.9%) compared with the 11–15-day treatment group (14.2%).
TABLE 1.
baseline demographics, clinical characteristics, and treatment characteristicsc
| 11–15-day therapy (N = 1161) |
5–10-day therapy (N = 246) |
Entirely intravenous (N = 349) |
Oral switch (N = 1112) |
|
|---|---|---|---|---|
| Demographics | ||||
| Age in years, median (IQR) | 68.2 (57.7–79.0) | 71.6 (60.7–80.9) | 67.9 (56.2–77.7) | 69.6 (59.3–80.1) |
| Sex, n (%) | ||||
| Female | 536 (46.2) | 106 (43.1) | 158 (45.3) | 508 (45.7) |
| Male | 625 (53.8) | 140 (56.9) | 191 (54.7) | 604 (54.3) |
| Race/ethnicity, n (%) | ||||
| Asian/Pacific Islander | 146 (12.6) | 30 (12.2) | 47 (13.5) | 134 (12.1) |
| Black/African American | 104 (9.0) | 22 (8.9) | 33 (9.5) | 96 (8.6) |
| Hispanic/Latino | 197 (17.0) | 44 (17.9) | 66 (18.9) | 180 (16.2) |
| White | 698 (60.1) | 144 (58.5) | 200 (57.3) | 683 (61.4) |
| Other/Unknown | 16 (1.4) | 6 (2.4) | 3 (0.9) | 19 (1.7) |
| Comorbidities | ||||
| Body mass index, median (IQR) | 29.0 (24.5–35.7) | 28.5 (24.3–35.1) | 29.7 (24.9–37.2) | 28.5 (24.2–35.0) |
| Charlson comorbidity index, n (%) |
||||
| CCI = 0 | 287 (24.7) | 55 (22.4) | 85 (24.4) | 263 (23.7) |
| 1 ≤ CCI ≤ 3 | 394 (33.9) | 83 (33.7) | 116 (33.2) | 384 (34.5) |
| CCI >3 | 480 (41.3) | 108 (43.9) | 148 (42.4) | 465 (41.8) |
| Dialysis dependence, n (%) | 24 (2.1) | 13 (5.3) | 5 (1.4) | 34 (3.1) |
| Diabetes mellitus, n (%) | 473 (40.7) | 98 (39.8) | 142 (40.7) | 449 (40.4) |
| Diagnosed with HIV, n (%) | 13 (1.1) | 4 (1.6) | 3 (0.9) | 15 (1.3) |
| Immunosuppressive therapy within preceding year | 325 (28.0) | 73 (29.7) | 96 (27.5) | 323 (29.1) |
| Concurrent diagnoses | ||||
| Cellulitis, n (%) | 347 (29.9) | 49 (19.9) | 108 (30.9) | 300 (27.0) |
| Cholangitis, n (%) | 18 (1.6) | 7 (2.8) | 6 (1.7) | 21 (1.9) |
| Cholecystitis, n (%) | 36 (3.1) | 8 (3.3) | 11 (3.2) | 33 (3.0) |
| Pneumonia, n (%) | 530 (45.7) | 133 (54.1) | 106 (30.4) | 587 (52.8) |
| Streptococcus spp., n (%) | ||||
| S. pneumoniae | 311 (26.8) | 71 (28.9) | 34 (9.7) | 356 (32.0) |
| GBS | 255 (22.0) | 42 (17.1) | 96 (27.5) | 210 (18.9) |
| S pyogenes | 165 (14.2) | 17 (6.9) | 62 (17.8) | 125 (11.2) |
| VGSa | 261 (22.5) | 85 (34.6) | 102 (29.2) | 263 (23.7) |
| Otherb | 169 (14.6) | 31 (12.6) | 55 (15.8) | 158 (14.2) |
| Hospital laboratory measures | ||||
| White blood cells, 109 cells/L, median (IQR) |
13.2 (8.9–17.9) | 13.2 (8.9–18.8) | 12.7 (8.9–17.8) | 13.3 (8.9–18.0) |
| Acuity level | ||||
| Admitted to ICU, n (%) | 214 (18.4) | 42 (17.1) | 64 (18.3) | 200 (18.0) |
| Discharge type, n (%) | ||||
| Home | 805 (69.3) | 170 (69.1) | 190 (54.4) | 818 (73.6) |
| Home with home health | 349 (30.1) | 73 (29.7) | 158 (45.3) | 285 (25.6) |
| Against medical advice | 7 (0.6) | 3 (1.2) | 1 (0.3) | 9 (0.8) |
| Antibiotic treatment | ||||
| Total antibiotic duration, median (IQR) |
15.0 (13.0–15.0) | 9.0 (8.0–10.0) | 15.0 (13.0–16.0) | 14.0 (11.0–15.0) |
| Antibiotic route at discharge, n (%) | ||||
| Intravenous | 302 (26.0) | 40 (16.3) | 349 (100) | 0 (0) |
| Oral | 859 (74.0) | 206 (83.7) | 0 (0) | 1112 (100) |
| Antibiotic group at discharge | ||||
| Oral beta lactam | 533 (45.9) | 99 (40.2) | 0 (0.0) | 659 (59.3) |
| IV beta lactam | 263 (22.7) | 35 (14.2) | 302 (86.5) | 0 (0.0) |
| Oral fluoroquinolone | 239 (20.6) | 62 (25.2) | 0 (0.0) | 316 (28.4) |
| Combination therapy | 91 (7.8) | 35 (14.2) | 39 (11.2) | 93 (8.4) |
| Other | 35 (3.0) | 15 (6.1) | 8 (2.3) | 44 (4.0) |
The VGS species category comprised isolates identified as viridans streptococci as well as S. anginosus, S. intermedius, S. milleri, S. gallolyticus, S. infantarius, S. mitis, S. sanguinis, S. salivarius, and S. mutans.
The “Other” species category comprised all other streptococcal isolates not included in the S. pneumoniae, GBS, S. pyogenes or VGS species categories and included those identified as group C streptococci, group G streptococci, and S. dysgalactiae.
Values are expressed as median (IQR) or n (%). GBS, Group B Streptococcus; HIV, human immunodeficiency virus; ICU, intensive care unit; IQR, interquartile range; spp., species; VGS, viridans group streptococci.
Treatment and disposition characteristics
The median duration of therapy was 14 days (IQR 11–15 days) overall, and 9 days (IQR 8–10 days) in the 5–10-day treatment group and 15 (IQR 13–15) in the 11–15-day treatment group (Table 1). There were differences in the route of antibiotic regimens on hospital discharge between groups, with patients in the 5–10-day treatment group more likely to be discharged on oral antibiotics compared with the 11–15-day treatment group (83.7% vs 74.0%). Antibiotic agents administered at discharge differed between groups, with more patients in the 5–10-day treatment group receiving oral fluoroquinolones, combination therapy, or other regimens versus the 11–15-day treatment group (25.2%, 14.2%, and 6.1% vs 20.6%, 7.8%, and 3.0%, respectively) and fewer patients in the 5–10-day treatment group receiving oral or intravenous beta-lactams (40.2% and 14.2% vs. 45.9 and 22.7%, respectively).
Outcomes
The overall cumulative incidence of the primary composite outcome was 20.8% and was not significantly different between groups (Table 2). The cumulative incidence of recurrent streptococcal bacteremia was 1.1%, which was slightly higher in the 11–15-day group (1.2%) compared with the 5–10-day group (0.8%, P < 0.001). Three of the 16 cases of recurrent streptococcal bacteremia died within the follow-up period, and two of these deaths seemed to be related to streptococcal infection. Hospital readmission occurred in 19.2% of patients overall, and differences were not detected between groups. The cumulative incidence of all-cause mortality in the 11–15-day group (4.5%) was higher than in the 5–10-day group (2.4%, P = 0.14). The overall cumulative incidence of C. difficile infection was 0.9%, and the incidence was lower (0%) in the 5–10-day treatment group than the 11–15-day treatment group (1.1%, P < 0.001).
TABLE 2.
Cumulative incidence of outcomesa
| Outcome, n (%) | 11–15-day therapy (N = 1161) |
5–10-day therapy (N = 246) |
Test of difference | Test of non-inferiority | Entirely intravenous (N = 349) |
Oral switch (N = 1112) |
Test of difference | Test of non-inferiority |
|---|---|---|---|---|---|---|---|---|
| Clostridioides difficile infection | 13 (1.1) | 0 (0.0) | <0.001 | <0.001 | 4 (1.1) | 11 (1.0) | <0.001 | <0.001 |
| Readmission | 221 (19.0) | 49 (19.9) | 0.75 | 0.080 | 75 (21.5) | 248 (22.3) | 0.75 | 0.04 |
| Recurrent streptococcal bacteremia | 14 (1.2) | 2 (0.8) | <0.001 | <0.001 | 2 (0.6) | 14 (1.3) | <0.001 | <0.001 |
| Death | 52 (4.5) | 6 (2.4) | 0.14 | <0.001 | 9 (2.6) | 59 (5.3) | 0.03 | 0.01 |
| Composite outcome | 241 (20.8) | 51 (20.7) | 0.99 | 0.047 | 80 (22.9) | 266 (23.9) | 0.70 | 0.06 |
Cumulative incidence expressed as N (%) for each treatment group.
In the unadjusted Farrington–Manning non-inferiority analysis, 5–10-day therapy was non-inferior to 11–15-day therapy with respect to the primary composite outcome (P = 0.047), death (P < 0.001), and recurrent streptococcal bacteremia (P < 0.001) (Table 2). With respect to hospital readmission, the 5–10-day therapy did not meet the threshold for non-inferiority (19.9% vs 19.0%, P = 0.08).
In the bivariate Cox proportional hazards regression model, 5–10-day treatment duration was not associated with the composite outcome (HR, 0.99; CI, 0.73–1.33, P = 0.97) (Table 3). Similarly, no association was found in the multivariable Cox proportional hazards regression model (aHR, 0.90; CI, 0.65–1.22, P = 0.50) (Table 3). Figure 2 shows the Kaplan–Meier survival curves.
TABLE 3.
Cox proportional hazards regression models for the primary composite outcome by treatment duration and routec
| Unadjusted HR | 95% CI | P-value | Adjusted HR | 95% CI | P-value | |
|---|---|---|---|---|---|---|
| 5–10-day treatment durationa | 0.99 | 0.73–1.33 | 0.97 | 0.90 | 0.65–1.22 | 0.50 |
| Oral switch antibiotic therapyb | 1.05 | 0.82–1.36 | 0.70 | 1.19 | 0.92–1.55 | 0.19 |
N = 1,407 for unadjusted analysis of antibiotic duration, and N = 1,390 for adjusted analysis of antibiotic duration. Patients who met the composite outcome within 10 days after initiating antibiotic therapy were excluded. In the adjusted analysis, 17 additional patients were excluded due to missing hospital white blood cell count.
N=1,461 for unadjusted analysis of antibiotic route, and N = 1,443 for adjusted analysis of antibiotic route. In the adjusted analysis, 18 additional patients were excluded due to missing hospital white blood cell count.
CI, confidence interval; HR, hazard ratio.
Fig 2.
Cumulative incidence of the primary outcome by duration of therapy. The Kaplan–Meier plot depicts the rate of the primary outcome among patients treated with 5–10 days versus 11–15 days of therapy.
Route of antibiotic therapy (i.e., oral switch vs. entirely intravenous)
Of the 1,461 patients included in the analysis, 1,112 patients received oral switch therapy, and 349 patients received entirely intravenous therapy.
Patient characteristics
Demographic and baseline characteristics are shown in Table 1. The median age was 69.3 years (IQR, 58.6–79.6), and the majority were men (54.4%). Patients in the oral switch group were more likely to have a concurrent diagnosis of pneumonia (52.8%) compared with the entirely intravenous group (30.4%). There were differences noted in the streptococcal species isolated, for example Streptococcus pneumoniae was more frequently isolated among patients in the oral switch group (32%) compared with the entirely intravenous group (9.7%).
Treatment and disposition characteristics
The median duration of therapy was 14 days (IQR, 11–15 days) in the oral switch group and 15 days (IQR 13–16 days) in the entirely intravenous group (Table 1). Fewer patients in the oral switch group were discharged with home health (25.6 vs 45.3%). Beta-lactam antibiotics were the most used class in both groups but accounted for a smaller portion of regimens in the oral switch group compared with the entirely intravenous group (59.3% vs 86.5%). Additionally, fluoroquinolones were commonly used in the oral switch group (28.4% of courses) but were not used in the entirely intravenous group.
Outcomes
The overall cumulative incidence for the primary composite outcome was 23.7%, for hospital readmission was 22.1%, and for C. difficile infection was 1.0%, and the rates did not meaningfully differ between groups (Table 2). The overall cumulative incidence of recurrent streptococcal bacteremia was 1.1% and was lower in the entirely intravenous group (0.6%) than in the oral switch group (1.3%, P < 0.001). The cumulative incidence of all-cause mortality was 4.7% overall, with a lower cumulative incidence in the entirely intravenous group (2.6%) than in the oral switch group (5.3%, P = 0.03).
We were unable to conclude non-inferiority in the unadjusted Farrington–Manning non-inferiority analysis of the primary composite outcome, with 22.9% of patients meeting the outcome in the entirely intravenous group vs 23.9% of patients in the oral switch group (P = 0.06) (Table 2). However, we found oral switch therapy was non-inferior to entirely intravenous therapy in the analysis of death (P = 0.01), recurrent streptococcal bacteremia (P < 0.001), and readmission (P = 0.04).
In the bivariate Cox proportional hazards regression model, oral switch therapy was not associated with the composite outcome (HR, 1.05; CI, 0.82–1.36, P 0.70) (Table 3). Similarly, no association was found in the multivariable Cox proportional hazards regression model (aHR, 1.19; CI, 0.92–1.55, P = 0.19) (Table 3). Figure 3 shows the Kaplan–Meier survival curves.
Fig 3.
Cumulative incidence of the primary outcome by route of therapy. The Kaplan–Meier plot depicts the rate of the primary outcome among patients treated with oral switch therapy versus entirely intravenous therapy.
DISCUSSION
The present study is the largest study to date to examine patient outcomes by both duration and route of therapy for uncomplicated streptococcal bacteremia. We found that patients treated with 5–10 days of antibiotic therapy did not have an increased rate of the primary composite outcome that included re-admission, recurrent streptococcal bacteremia, and death compared with those treated with 11–15 days of therapy. Although we were unable to conclude non-inferiority of oral switch therapy for the primary composite outcome, the incidence of the primary outcome was very similar between the oral switch and entirely intravenous groups. Additionally, in adjusted analyses, oral switch therapy was not associated with an increased risk of the primary outcome.
Our results align with prior studies examining the impact of treatment duration on outcomes. In a study by Nguyen et. al. that included 286 patients with S. pyogenes bacteremia, no difference was detected in 90-day mortality between patients treated with ≤10 days of antibiotic and those treated longer courses (24). Furthermore, a preliminary report by Gould et. al. that included 86 patients with non-staphylococcal gram-positive BSIs, 81% of which were streptococcal BSIs, determined that treatment outcomes were similar between patients treated with 6–10 days of therapy and those treated with 11–21 days of therapy (26). In contrast to the prior studies, ours was larger, and we included only streptococcal BSI of any species. Based on these prior findings and the finding of non-inferiority of 5–10-day courses in the present study, randomized clinical trials are warranted to confirm that shorter courses may be used preferentially to reduce the risks associated with prolonged courses.
Our results were also consistent with findings from multiple smaller retrospective studies, some of which are preliminary, that showed similar outcomes by route of therapy (28–32, 35, 36). Several groups found that all-cause mortality was similar among patients treated with oral switch therapy and those treated with entirely intravenous therapy (28, 30, 36). Others found no difference by route of therapy in various composite clinical outcomes (27, 29, 31, 32). All six groups evaluating impact on the length of hospital stay found that patients who transitioned to oral switch therapy had shorter hospitalizations (28–32, 36). Our study similarly did not find a statistically significant difference in the rate of the primary composite outcome, although the rate was numerically slightly higher in the oral switch group. We failed to conclude non-inferiority of oral switch therapy to entirely intravenous therapy possibly due to the slightly increased rate of the primary outcome and limited power as discussed below. Although outcomes seem relatively similar by treatment route in prior studies and the present study, and oral switch therapy has been successfully used in many cases, further investigation with a randomized controlled trial would be needed to definitively determine the optimal route of therapy on hospital discharge.
Our study had several limitations. The major limitation is that this is a retrospective study, so the threat of confounding by indication by duration and route of antibiotic therapy is present. To address this to the extent possible, we adjusted for potential baseline confounding in our multivariable models. Although cellulitis and pneumonia were not associated with the outcome in univariate analysis and were thus not included in our multivariable models, it is possible that the increased use of prolonged and intravenous therapy for cellulitis and short and oral switch therapy for pneumonia contributed to confounding by indication. Because some of the most ill patients receiving intravenous therapy may have been more likely to have been excluded for being discharged to a skilled nursing facility, it is possible that this selection bias contributed to the inability to conclude non-inferiority of oral regimens. Furthermore, the presentation of streptococcal bacteremia may vary by species (e.g., streptococcal toxic shock syndrome associated with S. pyogenes), but streptococcal species were considered together to power meaningful comparisons. We adjusted for streptococcal species in our multivariable models to mitigate potential confounding, but in choosing to analyze streptococcal species together, we would have been unable to detect any differences in response to treatment regimens by species.
Although our study was large, we had insufficient power for the non-inferiority analyses. We were able to include only 66% and 49% of the number of patients needed in the oral switch and entirely intravenous groups, respectively, to achieve 80% power as determined in our a priori power analysis. It is possible that the smaller than anticipated sample size contributed to the inability to conclude non-inferiority of oral switch therapy. In contrast, the use of a composite primary outcome may have introduced bias in favor of non-inferiority. Recurrent streptococcal bacteremia was the only component of the primary outcome specific to failure of streptococcal therapy. Because all-cause mortality and all-cause readmission occurred more frequently, it is possible that deaths and readmissions unrelated to the bacteremia could have biased results toward non-inferiority.
We did not account for the effect of antibiotic class or streptococcal sensitivity profile on patient outcomes. but anticipate both would have limited impact on our findings. Arensman et al. did not find a difference in comparing oral beta-lactam therapy to fluoroquinolone therapy (37). Additionally, the susceptibility data available from isolates tested in this study show that the majority of isolates were predicted to be susceptible to the antibiotics most commonly used.
A major strength of our study is the size, characterization, and follow-up of our study population that was drawn from a large integrated healthcare system. Furthermore, our results are likely to be generalizable, as the KPNC healthcare system provides care to approximately 34% of Californians. KPNC members are similar to the statewide population in terms of age, sex, race/ethnicity, but median household income was slightly higher in the KPNC population compared with the statewide population (38).
In summary, we found that clinical success rates in uncomplicated streptococcal bacteremia are similar when treated with 5–10-day courses versus 11–15-day courses and with oral switch therapy versus entirely intravenous therapy. We determined that 5–10-day courses were in fact non-inferior to longer courses and thus may be a safe and effective treatment option for uncomplicated streptococcal bacteremia. Further investigation with randomized controlled trials would be needed to confirm the safety and efficacy of shorter courses and to definitively determine the optimal route of therapy on hospital discharge.
ACKNOWLEDGMENTS
We thank Maqdooda Merchant of the KPNC Division of Research for her input on study design and statistical analysis.
D.S.C., Z.S.-Y., M.J.S., and J.H.C. received funding from a Kaiser Permanente Community Benefit Grant.
D.S.C. developed the study concept. D.S.C., Z.S.-Y., J.H.C., and M.J.S. developed the study design. J.S. and Z.S.-Y. contributed to feasibility analyses. Z.S.-Y. conducted all analyses. D.S.C. and J.H.C. performed data validation. D.S.C. and Z.S.-Y. wrote the manuscript. M.J.S. and J.H.C. edited the manuscript. All authors reviewed the final manuscript.
Contributor Information
Dana S. Clutter, Email: dana.s.clutter@kp.org.
Ryan K. Shields, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
ETHICS APPROVAL
This study was reviewed and approved by the KPNC Institutional Review Board. The requirement to obtain informed consent was waived.
SUPPLEMENTAL MATERIAL
The following material is available online at https://doi.org/10.1128/aac.00220-24.
Antibiotic susceptibility of streptococcal isolates.
ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.
REFERENCES
- 1. Laupland KB. 2013. Incidence of bloodstream infection: a review of population-based studies. Clin Microbiol Infect 19:492–500. doi: 10.1111/1469-0691.12144 [DOI] [PubMed] [Google Scholar]
- 2. Goto M, Al-Hasan MN. 2013. Overall burden of bloodstream infection and nosocomial bloodstream infection in North America and Europe. Clin Microbiol Infect 19:501–509. doi: 10.1111/1469-0691.12195 [DOI] [PubMed] [Google Scholar]
- 3. Rieger KL, Bosso JA, MacVane SH, Temple Z, Wahlquist A, Bohm N. 2017. Intravenous-only or intravenous transitioned to oral antimicrobials for Enterobacteriaceae-associated bacteremic urinary tract infection. Pharmacotherapy 37:1479–1483. doi: 10.1002/phar.2024 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Tamma PD, Conley AT, Cosgrove SE, Harris AD, Lautenbach E, Amoah J, Avdic E, Tolomeo P, Wise J, Subudhi S, Han JH, Antibacterial Resistance Leadership Group . 2019. Association of 30-day mortality with oral step-down vs continued intravenous therapy in patients hospitalized with Enterobacteriaceae bacteremia. JAMA Intern Med 179:316–323. doi: 10.1001/jamainternmed.2018.6226 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Daneman N, Shore K, Pinto R, Fowler R. 2011. Antibiotic treatment duration for bloodstream infections in critically ill patients: a national survey of Canadian infectious diseases and critical care specialists. Int J Antimicrob Agents 38:480–485. doi: 10.1016/j.ijantimicag.2011.07.016 [DOI] [PubMed] [Google Scholar]
- 6. Daneman N, Rishu AH, Xiong W, Bagshaw SM, Dodek P, Hall R, Kumar A, Lamontagne F, Lauzier F, Marshall J, Martin CM, McIntyre L, Muscedere J, Reynolds S, Stelfox HT, Cook DJ, Fowler RA, Canadian Critical Care Trials Group . 2016. Duration of antimicrobial treatment for bacteremia in Canadian critically ill patients. Crit Care Med 44:256–264. doi: 10.1097/CCM.0000000000001393 [DOI] [PubMed] [Google Scholar]
- 7. Alwan M, Davis JS, Daneman N, Fowler R, Shehabi Y, Rogers B. 2019. Duration of therapy recommended for bacteraemic illness varies widely amongst clinicians. Int J Antimicrob Agents 54:184–188. doi: 10.1016/j.ijantimicag.2019.05.011 [DOI] [PubMed] [Google Scholar]
- 8. Chotiprasitsakul D, Han JH, Cosgrove SE, Harris AD, Lautenbach E, Conley AT, Tolomeo P, Wise J, Tamma PD, Antibacterial Resistance Leadership Group . 2018. Comparing the outcomes of adults with Enterobacteriaceae bacteremia receiving short-course versus prolonged-course antibiotic therapy in a multicenter, propensity score-matched cohort. Clin Infect Dis 66:172–177. doi: 10.1093/cid/cix767 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Stevens V, Dumyati G, Fine LS, Fisher SG, van Wijngaarden E. 2011. Cumulative antibiotic exposures over time and the risk of Clostridium difficile infection. Clin Infect Dis 53:42–48. doi: 10.1093/cid/cir301 [DOI] [PubMed] [Google Scholar]
- 10. Bignardi GE. 1998. Risk factors for Clostridium difficile infection. J Hosp Infect 40:1–15. doi: 10.1016/s0195-6701(98)90019-6 [DOI] [PubMed] [Google Scholar]
- 11. Chastre J, Wolff M, Fagon J-Y, Chevret S, Thomas F, Wermert D, Clementi E, Gonzalez J, Jusserand D, Asfar P, Perrin D, Fieux F, Aubas S, PneumA Trial Group . 2003. Comparison of 8 vs 15 days of antibiotic therapy for ventilator-associated pneumonia in adults: a randomized trial. JAMA 290:2588–2598. doi: 10.1001/jama.290.19.2588 [DOI] [PubMed] [Google Scholar]
- 12. Vaughn VM, Flanders SA, Snyder A, Conlon A, Rogers MAM, Malani AN, McLaughlin E, Bloemers S, Srinivasan A, Nagel J, Kaatz S, Osterholzer D, Thyagarajan R, Hsaiky L, Chopra V, Gandhi TN. 2019. Excess antibiotic treatment duration and adverse events in patients hospitalized with pneumonia: a multihospital cohort study. Ann Intern Med 171:153–163. doi: 10.7326/M18-3640 [DOI] [PubMed] [Google Scholar]
- 13. Tamma PD, Avdic E, Li DX, Dzintars K, Cosgrove SE. 2017. Association of adverse events with antibiotic use in hospitalized patients. JAMA Intern Med 177:1308–1315. doi: 10.1001/jamainternmed.2017.1938 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Fallouh N, McGuirk HM, Flanders SA, Chopra V. 2015. Peripherally inserted central catheter-associated deep vein thrombosis: a narrative review. Am J Med 128:722–738. doi: 10.1016/j.amjmed.2015.01.027 [DOI] [PubMed] [Google Scholar]
- 15. Maki DG, Kluger DM, Crnich CJ. 2006. The risk of bloodstream infection in adults with different intravascular devices: a systematic review of 200 published prospective studies. Mayo Clin Proc 81:1159–1171. doi: 10.4065/81.9.1159 [DOI] [PubMed] [Google Scholar]
- 16. Herc E, Patel P, Washer LL, Conlon A, Flanders SA, Chopra V. 2017. A model to predict central-line-associated bloodstream infection among patients with peripherally inserted central catheters: the MPC score. Infect Control Hosp Epidemiol 38:1155–1166. doi: 10.1017/ice.2017.167 [DOI] [PubMed] [Google Scholar]
- 17. Krah NM, Bardsley T, Nelson R, Esquibel L, Crosby M, Byington CL, Pavia AT, Hersh AL. 2019. Economic burden of home antimicrobial therapy: OPAT versus oral therapy. Hosp Pediatr 9:234–240. doi: 10.1542/hpeds.2018-0193 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Bhagat H, Sikka MK, Sukerman ES, Makadia J, Lewis JS, Streifel AC. 2023. Evaluation of opportunities for oral antibiotic therapy in bone and joint infections. Ann Pharmacother 57:156–162. doi: 10.1177/10600280221101105 [DOI] [PubMed] [Google Scholar]
- 19. Iversen K, Ihlemann N, Gill SU, Madsen T, Elming H, Jensen KT, Bruun NE, Høfsten DE, Fursted K, Christensen JJ, Schultz M, Klein CF, Fosbøll EL, Rosenvinge F, Schønheyder HC, Køber L, Torp-Pedersen C, Helweg-Larsen J, Tønder N, Moser C, Bundgaard H. 2019. Partial oral versus intravenous antibiotic treatment of endocarditis. N Engl J Med 380:415–424. doi: 10.1056/NEJMoa1808312 [DOI] [PubMed] [Google Scholar]
- 20. Li H-K, Rombach I, Zambellas R, Walker AS, McNally MA, Atkins BL, Lipsky BA, Hughes HC, Bose D, Kümin M, et al. 2019. Oral versus intravenous antibiotics for bone and joint infection. N Engl J Med 380:425–436. doi: 10.1056/NEJMoa1710926 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Sawyer RG, Claridge JA, Nathens AB, Rotstein OD, Duane TM, Evans HL, Cook CH, O’Neill PJ, Mazuski JE, Askari R, et al. 2015. Trial of short-course antimicrobial therapy for intraabdominal infection. N Engl J Med 372:1996–2005. doi: 10.1056/NEJMoa1411162 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Yahav D, Franceschini E, Koppel F, Turjeman A, Babich T, Bitterman R, Neuberger A, Ghanem-Zoubi N, Santoro A, Eliakim-Raz N, Pertzov B, Steinmetz T, Stern A, Dickstein Y, Maroun E, Zayyad H, Bishara J, Alon D, Edel Y, Goldberg E, Venturelli C, Mussini C, Leibovici L, Paul M, Bacteremia Duration Study Group . 2019. Seven versus 14 days of antibiotic therapy for uncomplicated gram-negative bacteremia: a noninferiority randomized controlled trial. Clin Infect Dis 69:1091–1098. doi: 10.1093/cid/ciy1054 [DOI] [PubMed] [Google Scholar]
- 23. Kutob LF, Justo JA, Bookstaver PB, Kohn J, Albrecht H, Al-Hasan MN. 2016. Effectiveness of oral antibiotics for definitive therapy of Gram-negative bloodstream infections. Int J Antimicrob Agents 48:498–503. doi: 10.1016/j.ijantimicag.2016.07.013 [DOI] [PubMed] [Google Scholar]
- 24. Nguyen ADK, Smith S, Davis TJ, Yarwood T, Hanson J. 2023. The efficacy and safety of a shortened duration of antimicrobial therapy for group A Streptococcus bacteremia. Int J Infect Dis 128:11–19. doi: 10.1016/j.ijid.2022.12.015 [DOI] [PubMed] [Google Scholar]
- 25. Vu Y, Wong J, Liu C, Tverdek FP. 2022. 1819. Characterizing treatment duration of viridans group streptococci bacteremia in neutropenic cancer patients: is 14 days necessary? Open Forum Infect Dis 9. doi: 10.1093/ofid/ofac492.1449 [DOI] [Google Scholar]
- 26. Gould AP, Crawford RL. 2022. 1815. Assessment of therapy duration for non-staphylococcal gram-positive bloodstream infections. Open Forum Infect Dis 9. doi: 10.1093/ofid/ofac492.1445 [DOI] [Google Scholar]
- 27. Broermann L, Al-Hasan MN, Al-Hasan MN, Withers S, Benbow KL, Ramsey T, Ogren K, McTavish M, Webster W, Winders HR. 2022. 1843. Intravenous versus partial oral antibiotic therapy in the treatment of uncomplicated bloodstream infection due to Streptococcus species. Open Forum Infect Dis 9. doi: 10.1093/ofid/ofac492.1472 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Keintz MR, Torres CJ, Miller MM, Alexander BT, Lyden E, Lyden E, Lyden E, Ma J, Van Schooneveld TC, Marcelin JR. 2022. 1846. Outcomes in intravenous to oral antimicrobial therapy in beta-Streptococcus species. Open Forum Infect Dis 9. doi: 10.1093/ofid/ofac492.1475 [DOI] [Google Scholar]
- 29. Salam ME, Lew AK, Nguyen CT, Pettit NN, Pisano J. 2022. 1842. Intravenous to oral antibiotic stepdown for uncomplicated streptococcal bacteremia. Open Forum Infect Dis 9. doi: 10.1093/ofid/ofac492.1471 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Kang A, Bor C, Chen J, Gandawidjaja M, Minejima E. 2018. 1001. Evaluation of the clinical efficacy and safety of oral antibiotic therapy for Streptococcus spp. bloodstream infections. Open Forum Infect Dis 5:S297–S298. doi: 10.1093/ofid/ofy210.838 [DOI] [Google Scholar]
- 31. Ramos-Otero GP, Sarangarm P, Walraven C. 2022. A retrospective analysis of intravenous vs oral antibiotic step-down therapy for the treatment of uncomplicated streptococcal bloodstream infections. J Clin Pharmacol 62:1372–1378. doi: 10.1002/jcph.2097 [DOI] [PubMed] [Google Scholar]
- 32. Waked R, Craig WY, Mercuro NJ, Wungwattana M, Wood E, Rokas KE. 2023. Uncomplicated streptococcal bacteremia: the era of oral antibiotic step-down therapy? Int J Antimicrob Agents 61:106736. doi: 10.1016/j.ijantimicag.2023.106736 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Spellberg B, Talbot G, Infectious Diseases Society of America (IDSA), American College of Chest Physicians (ACCP), American Thoracic Society (ATS), Society of Critical Care Medicine (SCCM) . 2010. Recommended design features of future clinical trials of antibacterial agents for hospital-acquired bacterial pneumonia and ventilator-associated bacterial pneumonia. Clin Infect Dis 51:S150–S170. doi: 10.1086/653065 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Services USDoHaH, Administration FaD, Research CfDEa . 2018. Complicated urinary tract infections: developing drugs for treatment guidance for industry. Available from: https://www.fda.gov/media/71313/download
- 35. Ramirez JA, Bordon J. 2001. Early switch from intravenous to oral antibiotics in hospitalized patients with bacteremic community-acquired Streptococcus pneumoniae pneumonia. Arch Intern Med 161:848–850. doi: 10.1001/archinte.161.6.848 [DOI] [PubMed] [Google Scholar]
- 36. Loomis JM, Mang NS, Ortwine J, Prokesch BC, Wei W. 2022. Oral antibiotic stepdown therapy for uncomplicated streptococcal. Open Forum Infect Dis 9. doi: 10.1093/ofid/ofac492.028 [DOI] [Google Scholar]
- 37. Arensman K, Shields M, Beganovic M, Miller JL, LaChance E, Anderson M, Dela-Pena J. 2020. Fluoroquinolone versus beta-lactam oral step-down therapy for uncomplicated streptococcal bloodstream infections. Antimicrob Agents Chemother 64:e01515-20. doi: 10.1128/AAC.01515-20 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Gordon NP. 2020. Similarity of adult Kaiser Permanente members to the adult population in Kaiser Permanente’s northern California service area: comparisons based on the 2017/2018 cycle of the California health interview survey. Internal Kaiser Permanente report: unpublished. Kaiser Permanente [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Antibiotic susceptibility of streptococcal isolates.